Eating disorders and dissociation: A critical evaluation of the evidence

The papers identified explored various roles that dissociation can play within eating disorders. The following sections will firstly consider if there is an association between eating disorders and dissociation. It will then go on to explore the possible role of dissociation within eating disorders. These include the role of dissociation between trauma and eating disorders, self-harm and eating disorders, emotions and eating disorders, and dissociation in eating disorder symptomatology.

1.4.3.1. Association between eating disorders and dissociation

When considering the role of dissociation within eating disorders it is important to firstly consider if there is an association between dissociation and eating disorders. Four of the studies (Grave, Oliosi, et al., 1996; Grave et al., 1997; Schumaker et al., 1994;

Vanderlinden et al., 1995) explored the relationship between eating disorders and dissociation. Three of the studies found a significant difference in scores between individuals with an eating disorder diagnosis and control participants, with individuals with an eating disorder diagnosis reporting higher scores of dissociation. However, most of the studies tended to consider anorexia and bulimia when looking at dissociation and eating disorders and very few included individuals with Binge Eating Disorder and EDNOS.

On the other hand, most of these studies were conducted before the DSM 5 (American Psychiatric Association, 2013) was published so at that time Binge Eating Disorder would have been classified within EDNOS. All of these studies also recruited student samples as control participants. One study (Grave et al., 1997) considered dissociation within Binge Eating Disorders and found no significant difference between individuals with a diagnosis of Binge Eating Disorders who were obese and obese individuals without Binge Eating Disorders. Although no significant difference was found between these two groups, all participants were recruited from an inpatient unit and there were no non- clinical control participants, suggesting that there may be higher levels of dissociation within clinical samples of obese individuals. A further limitation of the present study was that it scored at the mid-point on the quality assessment framework due to a lack of a comprehensive review of the literature and missing information on the research process.

Three studies (Berger et al., 1994; Katz & Gleaves, 1996; McCallum et al., 1992) considered the prevalence of dissociative disorders within eating disorders and found that individuals with an eating disorder diagnosis and a comorbid dissociative disorder had significantly higher dissociative scores compared to individuals with an eating disorder diagnosis without a comorbid dissociative disorder. This suggests that dissociation within eating disorders may be part of a comorbid dissociative disorder rather than a distinct feature of eating disorders. For example, Katz and Gleaves (1996) recruited control participants and discovered that when compared to individuals with an

eating disorder diagnosis, they did not significantly differ on scores of dissociation. However, the present study had a very small sample size of fourteen participants per group and these findings should therefore be interpreted with caution.

In contrast, McCallum et al. (1992) found that individuals who had an eating disorder without a comorbid dissociative disorder reported significantly higher dissociative scores in relation to eating behaviour when compared to individuals with an eating disorder diagnosis and comorbid dissociative disorder. This suggests that dissociation may play a role within the symptomology of eating disorders. However, as dissociative disorders were diagnosed by means of clinical interview and inter-rater reliability was not investigated, there is some uncertainty regarding whether the comorbid diagnoses of dissociative disorders are accurate. Additionally, the present study scored at the mid- point on the quality assessment framework due to the method of data collection and lack of information on the method of data analysis, making it difficult to assess how significant these results are.

Overall, these studies suggest that there is an association between dissociation and eating disorders. However, consideration also needs to be given to co-morbid dissociative disorders. The following studies have explored the possible roles that dissociation may have within eating disorders.

1.4.3.2. The role of dissociation in trauma and eating disorders

One hypothesis on the role of dissociation in eating disorders is that it may act as a mediating factor between traumatic experiences and the eating disorder. Eight studies investigated the role of dissociation in the relationship between eating disorders and trauma but only one study explored if dissociation acts as a mediator. Everill, Waller and Macdonald (1995b) considered whether dissociation acts as a mediating factor specifically between sexual abuse and bulimic symptoms in eating disorders. They found that higher levels of dissociation were associated with greater frequency of

bingeing. There was no significant difference between individuals who reported experiences of abuse and those who did not in frequency of bingeing when dissociation was partialled out, suggesting that dissociation acts as a mediating factor between sexual abuse and the bingeing symptoms of eating disorders.

Restricting the study to a focus on sexual abuse only can be justified as four of the studies (Brown et al., 1999; Herzog et al., 1993; Léonard et al., 2003; Vanderlinden, Vandereycken, et al., 1993) have found that individuals with an eating disorder diagnosis who have experienced sexual abuse report higher levels of dissociation when compared to individuals diagnosed with eating disorders who have not experienced abuse. However, these results need to be interpreted with caution as three of the studies (Herzog et al., 1993; Léonard et al., 2003; Vanderlinden, Vandereycken, et al., 1993) had small sample sizes ranging from ten to thirteen participants who had experienced sexual abuse and had small samples for groups of individuals who had experienced other forms of abuse, which could account for the lack of significance in scores of dissociation. They also use self-report methods for assessing childhood abuse and one study (Vanderlinden, Vandereycken, et al., 1993) scored 18 on the quality assessment framework as it did not provide a clear rationale and used an unpublished measure to assess childhood trauma bringing to question the reliability and validity of the data collected.

In addition, Grave, Rigamonti, Todisco and Oliosi (1996) found no significant differences in scores of dissociation between individuals with an eating disorder diagnosis who have experienced sexual abuse as compared to those who have not. Similarly, Nagata, Kiriike, Iketani, Kawarada and Tanaka (1999) found no significant difference in dissociation scores for sexual abuse but they did find a significant difference when exploring physical abuse. Hartt and Waller (2002) found significant differences in scores of dissociation in individuals with an eating disorder diagnosis who have experienced neglect when compared to individuals with an eating disorder diagnosis who have not

experienced neglect. Once again these results need to be interpreted with caution as all three studies had small group sizes for the different categories of abuse. However, all of these studies scored above the mid-point on the quality assessment framework with two of the studies using validated measures of childhood trauma. One study (Hartt & Waller, 2002) scored 33 on the quality assessment framework, suggesting that it is important to consider the relation of other forms of abuse to eating disorders.

Everill et al. (1995b) also restricted their study to focus on individuals with a diagnosis of bulimia and more specifically looking at the role of dissociation in the bingeing symptoms of bulimia. This restriction can be seen as reasonable as three studies (Grave, Rigamonti, et al., 1996; Oliosi & Grave, 2003; Vanderlinden, Vandereycken, et al., 1993) have found high rates of experiences of abuse in individuals with a diagnosis of bulimia. Additionally, Léonard et al. (2003) found significantly higher scores of dissociation in individuals with a diagnosis of bulimia who have experienced abuse when compared to control participants. However, Hartt and Waller (2002) found no significant correlation between experiences of sexual abuse and bulimic symptomology in individuals with a diagnosis of bulimia and found a positive trend between trauma and vomiting behaviours, suggesting that trauma may be more strongly related to purging behaviours rather than bingeing behaviours. Nagata et al. (1999) also discovered no significant differences in experiences of abuse in individuals with bulimia when compared to control participants.

Comparing these studies along with the rates of abuse within eating disorders can be difficult as they have used a variety of methods to assess for history of abuse. In two studies (Everill et al., 1995b; Grave, Rigamonti, et al., 1996) researchers asked about experiences of abuse within the initial clinical interviews. Similarly, Vanderlinden et al. (1993) asked about experiences of abuse within a clinical interview and asked participants to complete a questionnaire. Five of the studies (Brown et al., 1999; Hartt & Waller., 2002; Herzog et al., 1993; Léonard et al., 2003; Nagata et al., 1999) utilised self-

report questionnaires. The use of these various methods of measuring trauma and abuse suggests that these studies employed different definitions of abuse which could have impacted on the reports of abuse. In addition, the use of self-report measures when assessing for experiences of abuse can result in under-reporting of such experiences by participants.

As there is only one study exploring the mediating role of dissociation in the relationship between eating disorders and trauma, it is unclear what role dissociation plays in this relationship. Additionally, as the study restricted the exploration to a specific type of abuse and a particular symptom of eating disorders there still remains uncertainty about the role that dissociation may play in the relationship between eating disorders and other forms of abuse.

1.4.3.3. The role of dissociation in eating disorders and self-harm

In considering the role of dissociation in the relationship between eating disorders and trauma, it has been suggested that dissociation may be a way of managing trauma and that self-harm is a way of coping with dissociative experiences. Claes and Vandereycken (2007) investigate these relationships using validated measures and found that there was a high probability of self-injurious behaviour in individuals with an eating disorder diagnosis who have had experiences of trauma. In particular, individuals who self-injured and experienced sexual abuse scored significantly higher on identity confusion, amnesia and absorption subscales of dissociation when compared to individuals who experienced sexual abuse and do not self-injure.

However, there were small numbers of participants who engaged in self-injurious behaviour who had experienced particular types of abuse. For example, there were only fourteen participants who experienced physical abuse and engaged in self-injurious behaviour. Additionally, there was no control group to compare these results and a large number of measures were utilised with no correction made to the probability in order to

avoid a type-1 error and if this correction was made it could cause their results to be insignificant. Muehlenkamp et al. (2011) investigated a conceptual model linking childhood trauma, eating disorders, dissociation and non-suicidal self-injury with a larger sample. They found a significant pathway from childhood abuse and low self-esteem to eating disorders and from eating disorders to dissociation to self-harm, suggesting that trauma and dissociation may make individuals with an eating disorder diagnosis more vulnerable to non-suicidal self-injury.

The hypothesis that self-harm may act as a way of coping with dissociative experiences has been investigated by one study (Noma et al., 2015). They found that dissociation was related to both recent and past incidents of non-suicidal self-injury, arguing that this may be due to self-injury being used as a mechanism for coping with dissociative states. Similarly, Demitrack et al. (1990) found that individuals with an eating disorder diagnosis who scored above the normal range on dissociation had a significantly higher frequency of self- harm and suicidal attempts when compared to individuals with an eating disorder diagnosis who scored within the normal range of dissociation. Two other studies (Claes et al., 2003; Paul et al., 2002) have also found significantly higher levels of dissociation in individuals with an eating disorder diagnosis who self-harm when compared to individuals who do not self-injure.

Demitrack et al. (1990) also found no significant difference in impulsivity or illness severity when comparing individuals with a diagnosis of eating disorders who engage in self-injurious behaviours with those who do not, suggesting that self-injurious behaviours may be a way of adapting to dissociative states. However, compared to the other studies, the present study scored lower on the quality assessment framework and did not use a validated measure for self-injurious behaviour.

In addition, Noma et al. (2015) found that dissociative scores were higher in individuals with an eating disorder diagnosis with a co-morbid disorder when compared to

individuals with an eating disorder diagnosis, suggesting that dissociation may be related to these co-morbid disorders. The three other studies (Claes et al., 2003; Demitrack et al., 1990; Paul et al., 2002) did not report on whether their sample had co-morbid disorders and this may have been a confounding variable. For example, Claes et al. (2003) reported that individuals with an eating disorder diagnosis who engaged in self- injurious behaviour reported higher levels of anxiety and depression.

These studies suggest that dissociation may play a role within the relationship of eating disorders and self-harm. However, consideration needs to be given to the impact of comorbid disorders and the role that anxiety and depression may have in this relationship.

1.4.3.4. The role of dissociation in the relationship between eating

disorders and emotion

When taking into consideration the impact of comorbid disorders in the relationship between eating disorders and dissociation Gleaves and Ebernez (1995b) argue that dissociation may not be a core symptom of eating disorders. They found severity of bulimic symptoms to be uncorrelated with dissociation and severity of anorexic symptoms were uncorrelated to dissociation once anxiety and depression were accounted for. However, due to the correlational nature of the study it cannot be interpreted to mean that dissociation does not play a role within eating disorders.

Two other studies (Greenes et al., 1993; Schumaker et al., 1995) have argued that dissociation may be related to symptoms of depression in individuals with bulimia. Greenes et al. (1993) discovered no differences in scores of dissociation between individuals with a diagnosis of bulimia and individuals with a diagnosis of depression but found that individuals with a diagnosis of bulimia and depressive symptoms scored higher then individuals with a diagnosis of bulimia and no depressive symptoms.

However, their sample sizes were small, with a range of seven to nine participants in each group, and they did not include a non-clinical control group within their study.

Schumaker et al. (1995) similarly explored the relationships between depression and dissociation in eating disorders, finding a moderate correlation between dissociation and depression in individuals with an eating disorder diagnosis which was not found in the control group. This suggests that dissociation and depression may be linked within eating disorders; however, it is important to consider that the control participants were a student population and there were no clinical controls. Additionally, both these studies scored nineteen on the quality assessment framework due to limited review of the literature, missing information on the methodology and limited information on the results of the study.

When exploring the role of dissociation in the relationship between eating disorders and comorbid psychological symptoms Farrington et al. (2002) found that scores of dissociation did not differ between adolescents with anorexia compared to a clinical and control sample. However, the mixed clinical group included individuals with a diagnosis of bulimia, which previous studies have argued can report significantly higher levels of dissociation (Demitrack et al., 1990; Schumaker et al., 1995; Waller et al., 2003). On the other hand, when exploring dissociation within individuals with a diagnosis of anorexia Farrington et al. (2002) discovered that scores of dissociation positively correlated with hostility and interpersonal sensitivity, suggesting that dissociation within Anorexia Nervosa is focused on the reduction of affect, particularly anger and interpersonal affect. Compared to the other studies, the present study also scored better on the quality assessment framework.

In an experimental study, Hallings-Pott et al. (2005) showed individuals with a diagnosis of any eating disorder involving binge eating and control participants lonely cues while observers measured levels of state dissociation. They found that, when shown lonely

cues, individuals with an eating disorder diagnosis did not experience an increase in anxiety or depression in relation to the cues but experienced significantly higher levels of state dissociation, in particular derealisation, when compared to neutral cues. There was no such difference in the non-clinical group highlighting how dissociation may play a role in managing interpersonal affect within eating disorders. However, the data was analysed using parametric analysis even though the data was not normally distributed and inter-rater reliability was not reported, suggesting that the findings need to be interpreted with caution.

Overall, when considering the role of dissociation in eating disorders it is important to consider the impact that other psychological symptoms may play in this relationship. However, it appears that in some way dissociation may play a role in managing particular affect within eating disorders.

1.4.3.5. The role of dissociation in eating disorder symptomatology

In addition to dissociation potentially being used as a strategy by individuals with an eating disorder diagnosis to reduce affect, it can also impact upon the symptoms within eating disorders. Dissociation has been considered to play a role within the bingeing symptoms of eating disorders and two studies (Everill et al., 1995a; La Mela et al., 2010) have found significant correlations between bingeing behaviour and scores of dissociation. However, frequency of bingeing episodes in these studies may not be accurate as self-report methods were used to identify the number of bingeing episodes.

Everill et al. (1995a) conducted a further multiple regression analysis and found dissociation accounted for 39% of the variance in bingeing frequencies, suggesting that dissociation plays a specific role in bingeing behaviour. Similarly, Lyubomirsky, Sousa and Casper (2001) found that individuals with an eating disorder diagnosis reported experiences of dissociation during and after bingeing when compared to a non-clinical

sample who engage in occasional bingeing behaviour but these results need to be considered with caution as no standardised measure of dissociation was used.

Engelberg et al. (2007) employed an ecological momentary assessment paradigm to monitor dissociation prior to and during bingeing episodes. They found that the presence of dissociation prior to bingeing increased the probability that individuals would engage in bingeing behaviour. This suggests that dissociation may play a role in the onset of bingeing behaviour. However, as dissociation was measured with the use of only three adapted questions it is difficult to identify whether the experiences reported prior to a binge were dissociative in nature. In contrast to these studies, Waller et al. (2003) found significantly higher levels of psychological dissociation in individual with an eating disorder who engage in purging behaviour and significantly higher levels of somatoform and psychological dissociation in relation to excessive exercise, suggesting that